Reliability analysis of portal frame subjected to varied lateral loads using machine learning: Fid-Bau Portal

Reliability analysis of portal frame subjected to varied lateral loads using machine learning

Sufyan, Md Saeb / Samui, Pijush / Mishra, Shambhu Sharan

Structural reliability analysis has a vital significance in assessing the performance and safety of engineering structures. Traditional methods of reliability analysis often rely on deterministic models, which may not accurately represent the uncertainties and variability present in real-world scenarios. To address this limitation, the paper proposes the integration of machine learning techniques to enhance the accuracy and efficiency of reliability analysis for portal frames subjected to lateral loads. Portal frames are commonly used in low-rise buildings that require open and flexible spaces, such as industrial, commercial, agricultural, and storage facilities. The goal of this study is to find a reliable and efficient method that can accurately predict the structural response and failure probability of portal frames subjected to lateral forces. To estimate the displacement subjected to lateral forces, three hybrid random forest (RF) models have been developed in this study: random forest-dragonfly optimization algorithm (RF-DOA), random forest-sparrow search algorithm (RF-SSA) and random forest-whale optimization algorithm (RF-WOA). The displacement due to lateral forces has been efficiently predicted by all the proposed models. When the three models were compared, the RF-WOA showed best prediction and high accuracy during the testing phase. Also, the RF-WOA model outperformed the RF-DOA and RF-SSA algorithms, based on the outcomes of rank analysis, regression line analysis, and reliability analysis. Therefore, the RF-WOA model can be used as the most precise machine-learning algorithm to determine displacement subjected to lateral forces. Engineers and designers can benefit from the developed methodology in optimizing the analysis and design of portal frames, thereby enhancing the safety of structures.